Human Action Recognition (HAR) aims to understand human behavior and assign a label to
each action. It has a wide range of applications, and therefore has been attracting increasing …

Dynamic neural network is an emerging research topic in deep learning. Compared to static
models which have fixed computational graphs and parameters at the inference stage …

While state-of-the-art vision transformer models achieve promising results in image
classification, they are computationally expensive and require many GFLOPs. Although the …

Effective spatio-temporal modeling as a core of video representation learning is challenged
by complex scale variations in spatio-temporal cues in videos, especially different visual …

Streaming video clips with large-scale video tokens impede vision transformers (ViTs) for
efficient recognition, especially in video action detection where sufficient spatiotemporal …

In this paper a pure-attention bottom-up approach, called ViGAT, that utilizes an object
detector together with a Vision Transformer (ViT) backbone network to derive object and …

Prototypical network is a key algorithm to solve few-shot problems. Previous prototypical
network based methods average sentence embeddings of the same class to obtain …

This paper introduces a new and challenging Hidden Intention Discovery (HID) task. Unlike
existing intention recognition tasks, which are based on obvious visual representations to …

Action detection plays an important role in high-level video understanding and media
interpretation. Many existing studies fulfill this spatio-temporal localization by modeling the …

Abstract Contrastive Language-Image Pretraining (CLIP) models have achieved significant
success and have markedly improved the performance of various downstream tasks …